Files
rmancinas 7e65fc81d5 Add legacy database structural reference for non-Windows machines
docs/LEGACY_DATABASES.md documents all three source Access databases
(every table, column, type, and known data-quality quirk) generated
from a live read of the real files, so no Windows/Access driver is
needed to understand their structure going forward.

New migration/ tooling: catalog_schema.py connects to the real files
and walks every table (including excluded scratch tables);
render_catalog_md.py renders that into the doc's appendix. Raw output
checked in at migration/catalog.json so the doc can be regenerated
without touching Access again.
2026-07-22 01:41:53 -07:00

109 lines
3.6 KiB
Python

"""
Generates a complete structural catalog of every table in the legacy Access
source files (including excluded/scratch tables, for documentation
completeness) and writes it as JSON.
This is the source data for docs/LEGACY_DATABASES.md. Re-run this any time
the source .accdb/.mdb files change, then regenerate the doc from the JSON.
Must run on a Windows machine with the Microsoft Access Database Engine
ODBC driver installed (pyodbc can't reach Access files otherwise) — see
docs/LEGACY_DATABASES.md for why non-Windows machines can't do this step
themselves and have to work from this catalog / the staged Parquet output
instead.
Usage:
python catalog_schema.py --output catalog.json
"""
from __future__ import annotations
import argparse
import json
from pathlib import Path
import extract
from config import SOURCES
def get_access_type_labels(cnxn, table_name: str) -> dict[str, str] | None:
"""Best-effort: cursor.columns() gives real Access type names/sizes
(VARCHAR(255), DOUBLE, LONGCHAR, LONGBINARY, COUNTER, BIT, CURRENCY...)
but decode-errors on a subset of tables (see extract.py). Returns None
on failure so the caller can fall back to pandas-inferred dtypes."""
try:
cursor = cnxn.cursor()
labels = {}
for col in cursor.columns(table=table_name):
size = col.column_size
labels[col.column_name] = f"{col.type_name}({size})" if size else col.type_name
return labels
except Exception:
return None
def catalog_source(source_name: str, source_cfg: dict) -> dict:
path = source_cfg["path"]
result = {"path": str(path), "tables": {}}
if not Path(path).exists():
result["error"] = "file not found"
return result
cnxn = extract.connect(path)
tables = extract.list_tables(cnxn)
for table_name in tables:
entry = {"excluded": table_name in source_cfg["exclude"]}
access_types = get_access_type_labels(cnxn, table_name)
try:
df = extract.read_table(cnxn, table_name)
except Exception as exc: # noqa: BLE001
entry["error"] = str(exc)
result["tables"][table_name] = entry
continue
original_columns = df.attrs.get("original_columns", list(df.columns))
columns = []
for orig_name, sanitized_name in zip(original_columns, df.columns):
dtype = str(df[sanitized_name].dtype)
has_nulls = bool(df[sanitized_name].isna().any()) if len(df) else None
type_label = None
if access_types is not None:
type_label = access_types.get(orig_name)
columns.append(
{
"name": orig_name,
"access_type": type_label,
"pandas_dtype": dtype,
"nullable": has_nulls,
}
)
entry["row_count"] = len(df)
entry["column_count"] = len(columns)
entry["columns"] = columns
entry["access_type_metadata_available"] = access_types is not None
result["tables"][table_name] = entry
return result
def main() -> None:
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("--output", type=Path, default=Path("catalog.json"))
args = parser.parse_args()
catalog = {}
for source_name, source_cfg in SOURCES.items():
print(f"Cataloging {source_name}...")
catalog[source_name] = catalog_source(source_name, source_cfg)
args.output.write_text(json.dumps(catalog, indent=2, default=str), encoding="utf-8")
print(f"Wrote {args.output}")
if __name__ == "__main__":
main()